Earlier today I noticed constant issues for connections from Primary and Worker to Redis in a stable n8n deployment. I updated all deployments but the connection remained unstable. I tried to replace public host and port by their private counterparts but then faced apparently DNS issues with connections failing. I had to revert to the public proxy. Still unstable. How to fix this?

ECONNRESETs every ~30s through roundhouse proxy starting today at 9:45 AM, Redis itself shows zero load.